Symptomatic knee osteoarthritis has an incidence of 240/100,000 person years and is one of the most frequent causes of dependency in lower limb tasks, especially in the elderly. It causes 68 million work loss days per year and more than 5% of the annual retirement rate. Osteoarthritis is the most frequent reason for joint replacement at a cost to the community of billions of dollars per year. No effective medical remedies for osteoarthritis currently exist. However, the pharmaceutical industry is attempting to develop drugs that retard progression of OA. If efficacious, these proprietary medications will be expensive to employ in a population in which OA is endemic. Ironically, there is evidence that vitamin D supplementation, a simple non-proprietary intervention, may have efficacy in slowing progression of OA. Even if only modestly effective, it could have considerable impact in terms of reducing the societal burden of OA. Therefore, in the interests of public health, the efficacy of vitamin D supplementation as a disease-modifying treatment for OA needs to be tested in a rigorous clinical trial. Disease-modification trials for knee OA have been difficult due to limitations of the radiographic technique. However, MRI has emerged as a valid, precise and reproducible tool for obtaining volumetric measures of cartilage and joint structures. Our goal is to enroll 140 individuals with symptomatic knee OA into a 2-year randomized placebo controlled clinical trial of vitamin D, 14,000 ID / week. Outcomes will include the WOMAC questionnaire (primary clinical), cartilage volume loss (primary pathological), physical function tests, SF-36 and whole organ knee MRI OA severity scores. Bone density, calcium homeostasis and knee alignment will be tested as intermediary variables. We will test effectiveness of vitamin D using cross-sectional time series regression models with first-order autoregressive disturbance adjusting for informative dropouts.